Desalination 222 (2008) 66–73

UF/MF pre-treatment to RO in seawater and wastewater

reuse applications: a comparison of energy costs

G.K. Pearce
Membrane Consultancy Associates, P.O. Box 4006, Pangbourne RG8 7WB, UK
Tel. +44 118 9843106; Fax +44 118 9845957; email: gpearce@membraneconsultancy.com
Received 6 May 2007; accepted 17 May 2007

Abstract
Water resources are becoming increasingly scarce in many areas of the world due to development and increased
demand. In consequence, the market for reverse osmosis (RO) is expanding to meet the increasing requirement by
use of seawater and wastewater resources. Membrane filtration has gradually gained acceptance as the preferred
pre-treatment to RO. However, although perceived as desirable, UF/MF is also thought to be an expensive option,
and consideration of UF/MF is sometimes restricted to applications which are thought to be especially problematic.
In wastewater treatment applications, UF/MF is the pretreatment technology of choice due to the highly fouling
nature of the feed. This paper provides examples of the energy cost for various water sources, comparing waste-
water reuse with surface water, brackish water and seawater desalination. In the wastewater case, conventional
activated sludge followed by UF/MF-RO is compared with MBR-RO. The comparison shows that wastewater
reuse is a very attractive energy option, and that schemes should be considered where possible using UF/MF-RO
after conventional sewage treatment, or MBR for smaller schemes, or where space is at a premium.
Conventional surface water sources have an energy cost of 0.1–0.3 kW h/m3, with brackish water sources
normally falling in the range 0.8–1.7 kW h/m3. The energy cost of wastewater reuse from conventional treatment
is in the range 0.8–1.2 kW h/m3, whilst MBR-RO is in the range 1.2–1.5 kW h/m3. Seawater has the most expensive
energy cost, with a medium salinity feed needing an energy of 2.3–4.0 kW h/m3. Although the non-conventional
sources all use higher energy than surface or groundwater sources, the cost of transfer and distribution should also
be considered. A typical power cost for distribution is 0.6 kW h/m3, so proximity of the source and demand may
be a key factor. The data can be used to provide guidance for resource development.
Keywords: Ultrafiltration; Microfiltration; Reverse osmosis; Desalination; Seawater; Pre-treatment; Wastewater;
Reuse; Membrane bioreactor; Total water cost

1. Background demand. In consequence, the market for reverse

Water resources are becoming increasingly osmosis (RO) is expanding to meet the increasing
scarce in many areas of the world due to develop- requirement by use of seawater and wastewater
ment, particularly in costal regions, and increased resources. The relative contribution of various

Presented at the conference on Desalination and the Environment. Sponsored by the European Desalination Society
and Center for Research and Technology Hellas (CERTH), Sani Resort, Halkidiki, Greece, April 22–25, 2007.

0011-9164/06/$– See front matter © 2006 Published by Elsevier B.V.

doi:10.1016/j.desal.2007.05.029
 G.K. Pearce / Desalination 222 (2008) 66–73
Table 1
Water source as a contribution to total water abstraction
Flow
mLd
Surface water 6,525,000
Groundwater 3,751,000
Desalination 30,410
Water reuse 19,450
Total abstraction 10,326,000
water sources to the total water abstracted for all
uses is illustrated in Table 1 [1].
Water supply is dominated by surface water
sources, which comprise more than 60% of the
total. Most of the remainder is provided from
groundwater supplies. Both desalination and water
reuse are relatively small sources at present.
However, groundwater is effectively a non renew-
able resource, so it is inevitable that the contri-
bution of groundwater to the total abstraction
will diminish.
On the other hand, water demand is increasing
due to the following factors:
• Increasing population, and migration to
drought prone regions
• Industrial development, and increasing water
use per capita
• Climate change leading to changing weather
patterns in populated areas
Desalination and water reuse are therefore
predicted to be moving into a period of signifi-
cant sustained growth to compensate diminishing
supply and increased demand.
Desalination is now a mature technology that
has been actively developed over the past 35 years.
Though well established, it provides a relatively
small fraction of total water abstraction at just
under 0.3%, but growing at a significant rate of
approximately 8% p.a. Its great advantage is that
it often provides a water resource near where it is
needed, since a high proportion of the world’s
67
population lives near the coast, especially in
water stressed areas. The major disadvantage of
desalination is that it is energy intensive.
Water reuse is a relatively new market, though
% growing rapidly, and already supplying just
under 0.2% of total water abstraction. With a fore-
63.2 cast annual growth rate of 14%, it is predicted to
36.3
outstrip desalination by 2020. Standard treat-
0.29
0.19 ment for wastewater normally comprises a bio-
logical treatment stage, known as conventional
activated sludge (CAS), and clarification. This
degree of treatment is known as secondary treat-
ment. If followed by filtration, e.g. by a sand filter,
the treatment is known as tertiary treatment.
Historically, 70% of reused wastewater has only
been treated to a secondary or tertiary standard,
which would only be suitable for agricultural
use in less developed parts of the world. To be
considered for re-introduction to the drinking
water supply chain, and for most industrial uses,
wastewater normally requires a further level of
treatment, i.e. a quaternary stage, e.g. by RO/NF,
ion exchange or electro dialysis (EDI).
With the widespread acceptance of ultrafiltra-
tion (UF) and microfiltration (MF) technology
in the last 10–15 years, quaternary treatment has
become by far the most rapidly growing segment
of the wastewater reuse market. Indeed, when
membrane filtration technology is combined with
the biological reaction stage, a new unit operation
has become possible, i.e. the membrane bioreactor
(MBR). UF/MF may not require tertiary treat-
ment by sand filters as a pre-treatment, so the
term quaternary treatment becomes a misnomer.
Indeed, MBR does not need any pre-treatment
other than screening, and produces a feed suit-
able for RO/NF in a single treatment step. Table 2
illustrates the extent of wastewater treatment in
terms of flowrate.
Prior to the introduction of membrane filtra-
tion, the application of RO/NF in wastewater reuse
was restricted due to fouling problems. However,
UF/MF and MBR provides an excellent feed
quality for further treatment, and this technological
68 G.K. Pearce / Desalination 222 (2008) 66–73

Table 2 However, the RO industry has been sur-

Wastewater flows as a function of treatment standard prisingly slow to focus attention on the issue of
pre-treatment, and, with the exception of China,
Flow (mLd) often relies on conventional technology devel-
WW collected 1,011,000 oped for different requirements. To get the most
WW treated 438,000 out of the latest membrane and system design
WW reuse 19,450 developments, pre-treatment needs to provide a
WW treated to 4ry std 5500 better product quality for the RO feed, and made
WW treated by RO/NF 4000 more consistent and reliable.
Pre-treatment is the bugbear of the RO indus-
try. Whilst conventional pre-treatment technology
advance, combined with the market require- can be effective, it needs to be carefully designed,
ments, has led to the rapid rise in wastewater and diligently operated. Upsets, due to feed
reuse schemes. variability or contamination, will be transferred
Table 3 shows the segmentation of the waste- to the RO, sometimes with dire consequences.
water treatment market by end use [1,2]. Most examples of RO system failure can be put
RO has emerged as the most suitable tech- down to pre-treatment failings, either in design or
nology for addressing water needs in most areas, operation. The cost to rectify these shortcomings,
since it is a flexible cost effective technology and the lost production which results, threatens
with a mainly good track record. to give RO a bad name in some cases.
Two important trends have emerged in the This paper reviews the case for ultrafiltration
last 15 years of RO development. Firstly, RO (UF) and microfiltration (MF) pre-treatment to
membrane performance has improved markedly, RO. For brackish water and wastewater feeds,
and secondly, prices have reduced sharply as the case for UF/MF pre-treatment is normally
markets have expanded and projects have become accepted. The world’s largest industrial market
larger. Prior to this, new resource development for RO is in China, and there, UF/MF is specified
was normally several times cheaper than recov- for a significant proportion of RO projects irre-
ering a saline source with RO; now, the RO spective of feed type. Wastewater projects also
option is often cost competitive, and provides an normally use UF/MF, since the fouling nature of
independent flexible option to a project devel- the feed is recognized, and UF/MF significantly
oper. Recently significant improvements have improves the pre-treatment. In both cases, UF/MF
been made in system design and energy recovery, ensures that significantly higher fluxes are obtained
enhancing the RO option even further. from the RO, with much less fouling, reduced
chemical usage, and better on-stream time.
However, the case for seawater feeds is much
Table 3 less clear cut, especially for high salinity feeds,
Wastewater reuse market segmentation by end user since fluxes cannot be increased much due to
category osmotic pressure limitations. Also, the additional
expense of UF/MF is assumed to be too high,
%
making the project unattractive. The paper will
Industry and power 65.5 review the case for UF/MF pre-treatment, and
Municipal 11 compare the energy costs of seawater desalination
Agriculture 23 with surface and brackish water sources, and with
Other 0.5 wastewater reuse. For the reuse case, membrane
 G.K. Pearce / Desalination 222 (2008) 66–73
treatment after conventional activated sludge
will be compared to MBR. The comparison will
provide guidelines for resource development.
2. Advantages of membrane pre-treatment
to RO
The objective of pretreatment to an RO or NF
system is to remove particles, reduce organics,
and provide a feed that will not cause biofouling in
the RO/NF elements. Conventional pre-treatment
technology relies on a combination of chemical
treatment and media filtration to achieve condi-
tioning of the feed to make it acceptable as an
RO/NF feed by removing a proportion of the feed
challenge. In contrast, UF/MF uses a sieving
mechanism, which provides an absolute barrier to
particles above the size of the UF/MF membrane
pores, and thus can provide a much better RO feed.
Dissolved species such as salts and organic
solutes, pass through the UF/MF membrane.
Organics may be a problem to the RO/NF, since
they can cause fouling due to surface adsorption,
or they may provide a food source to micro-
organisms. It is therefore necessary that the effect
of these organics on the RO is investigated, and
that methods are investigated to remove them, or
to mitigate their effect.
Since membranes provide a barrier to par-
ticulates, they provide significant benefits to the
RO/NF, and to the overall system design. These
benefits fall into two broad categories, namely
reduced cost of the overall system, and improved
on-stream time and security of supply. The
benefits can be summarized as follows:
• Significantly higher RO design flux and
recovery is normally possible
• Low space – >33% saving with UF/MF
• RO membrane replacement rate reduced
significantly
• Can treat surface water, with poor and/or
variable quality
• Reduced requirement for RO disinfection and
cleaning
69
Membrane filtration is accepted as the best
pre-treatment alternative for some feed sources
and in some markets, whereas it is considered
expensive or unnecessary in other cases. For
example, membranes are accepted in all markets
for wastewater pre-treatment, whilst for surface
water, acceptance varies with location. For sea-
water, membrane pre-treatment is considered
expensive, and is only just beginning to gain
acceptance.
The status of membrane filtration for RO
pre-treatment for various source waters is sum-
marized below:
Surface water – membrane filtration accepted;
uptake greater in new markets
– UF/MF provides clear RO performance
improvement
– Improves stability and reliability
– Reduces total water cost
– Often specified for the industrial expansion
in China
Wastewater – membrane filtration accepted
– UF/MF provides much better RO feed quality
than conventional
– Removes particles and organics due to use of
coagulant
– Used as standard design for new systems
– Significant potential for MBR-RO
Seawater – membrane filtration case emerging
– Considered expensive
– Important advantages for open intake
– Improves opex, total water cost, and plant
on-stream time
As shown above, membranes are used in
wastewater pre-treatment primarily for technical
reasons, whereas for other sources, both membrane
and conventional pre-treatment can perform the
duty and the design is decided commercially on
a case by case basis.
70 G.K. Pearce / Desalination 222 (2008) 66–73
Table 4
Energy costs of treatment for surface and groundwater
sources
Power
(kW h/m3)
Surface water – conventional treatment 0.15–0.30
Surface water – conventional + UF/MF 0.25–0.35
Groundwater infl. by surface – UF/MF 0.1
3. Energy costs from conventional sources
Typical energy costs of producing potable
water from surface and groundwater sources are
shown in Table 4. The tabulated costs cover the
power cost of the treatment works itself, and are
dependent upon the extent of treatment require-
ments. If UF or MF polishing is used after con-
ventional treatment, the additional power cost is
low at about 0.1 kW h/m3, which is similar to
the requirement for groundwater schemes. If
UF/MF is used in place of conventional processes
for treating raw surface water, power costs will
be higher than for polishing and will depend on
pre-treatment requirements, but are likely to be
around 0.2 kW h/m3. A separate entry has not
been included in the table for this type of treatment
since in many cases, such treatment may not
meet all of the treated water quality parameters,
and therefore is not strictly comparable with
conventional treatment.
In some parts of the world, coastal aquifers
are brackish due to saline intrusion. Table 5
summarizes the costs of treating brackish sources
Table 5
Brackish water power costs from three case studies
Feed Salinity Pre-treatment
TDS (ppm) (kW h/m3)
Brackish surface water 930 0.2
Brackish groundwater 2200 0.1
Estuary case study 3000–11,000 0.29
from three typical cases. The groundwater exam-
ple is a relatively high salinity aquifer based on
a case study in Florida, in which pre-treatment
requirements are low, whilst the surface water
is a low salinity river water with greater pre-
treatment requirements. These two cases are taken
from the large element consortium study [3]. The
estuary case is typical of many schemes currently
being considered in which there is a significant
tidal influence on feed salinity, partially abated
by management of abstraction timing.
The power costs of brackish water treatment
are significantly greater than from conventional
sources due to the power requirements of the
RO in overcoming osmotic pressure. However,
the impact of the RO is not as great as may have
been expected.
An important factor to consider is the cost of
distribution. A typical figure for distribution costs
in the UK is 0.6 kW h/m3, which is significant in
comparison to the costs of the actual treatment,
and indeed exceeds treatment costs for the conven-
tional sources considered in Table 4. It is there-
fore evident that it is important to secure water
sources and provide treatment as close as possible
to where it is needed to control transfer costs.
4. Energy cost for wastewater reuse
Treatment of wastewater by a biological pro-
cess followed by clarification is typically known
as the conventional activated sludge process
(CAS), and produces a secondary treated effluent.
This is not suitable as an RO feed, and requires
RO system Total treatment Distribution
(kW h/m3) (kW h/m3) (kW h/m3)
0.63 0.83
0.88 0.98
1.38 1.67 2.27
 G.K. Pearce / Desalination 222 (2008) 66–73 71

Table 6 evident that the RO also suffers from very high

2ry/3ry wastewater treatment; the O&M cost of pre- operating costs if conventional pre-treatment is
treatment to an RO feed standard (MWH, West Basin, used since both power and cleaning costs for the
based on a 19 mLd facility) RO stage are significantly greater than if mem-
brane filtration is used.
Conv. MF
The data are from Orange County [4,5] for a
3
Utilities (kW h/m ) 51 mLd scale facility, and show that conventional
Power 0.2 0.2 pre-treatment has a major effect of increasing
3
O&M costs (US cents/m ) energy cost in the RO, due to fouling and high
Power 2 1.8 operating pressures. Since membrane filtration
Other 14 3.2 has lower capex and other operating costs than
Total 13 5 conventional pre-treatment, it is now widely
accepted that membrane pre-treatment is man-
further treatment. Prior to the introduction of datory for any reuse scheme. The membrane
UF/MF and MBR, conventional treatment stages stage can either be based on UF/MF with a CAS
of lime clarification and sand filtration were used. feed, or on MBR. The energy cost of the RO
However conventional pre-treatment has very stage combined with MF is 0.5 kW h/m3, indicat-
high chemical costs, and so is unattractive for RO ing that the RO energy cost alone would be just
pre-treatment. Table 6 shows the comparison of 0.4 kW h/m3, i.e. significantly below the levels
conventional and membrane filtration in treating found in seawater or medium salinity brackish
secondary or tertiary wastewater to an RO feed water applications.
standard [4]. Metcalfe and Eddy carried out a survey of
It is notable that whilst power costs for con- wastewater facilities in the US [6], and found that
ventional and membrane filtration pre-treatment the energy usage range was 0.32–0.66 kW h/m3.
are very similar, conventional pre-treatment has Energy usage in wastewater treatment is some-
much higher operating costs overall, since chemi- what lower in Europe according to Black and
cal dosing costs for this type of treatment are Veatch (B&V) [7], who have carried out extensive
very high. surveys of wastewater treatment costs. This is
Furthermore, Table 7 shows the effect on the partly due to a greater consciousness for energy
RO of using membrane filtration as pre-treatment efficiency, and partly due to the fact that the aver-
compared to conventional pre-treatment [4]. It is age BOD loading/capita in Europe is typically
70% of that in the US (due to the use of kitchen
disposal units in the US). Also, some of the
Table 7 facilities in the US survey were old and ineffi-
Energy comparison of RO system requirements
cient in energy usage, whilst those surveyed in
including pre-treatment
Europe were relatively modern.
Conv. + RO MF + RO
European wastewater costs can be as low as
about 50% of those of the US. Long term moni-
Utilities (kW h/m3) toring of a wastewater treatment system in
Power 1.3 0.5 Denmark showed a usage of 0.15 kW h/m3 for
3
O&M costs (US cents/m ) activated sludge, increasing to 0.25 kW h/m3 if a
Power 13.2 5.3 BAF stage is included. Monitoring of a German
Other 12.7 10.6 system confirmed a usage of 0.25 kW h/m3 for a
Total 25.9 15.9
similar comprehensive treatment system of CAS
72 G.K. Pearce / Desalination 222 (2008) 66–73
Table 8
Typical MBR energy costs
Municipal sewage
(kW h/m3)
Microbiology 0.3
MBR-air for fouling control 0.4
and BAF. The energy cost of membrane filtra-
tion according to B&V is 0.05–0.10 kW h/m3,
which is broadly in agreement with the data pre-
sented in the tables above. RO has an energy cost
of 0.25–0.50 kW h/m3, which is also consistent.
Assessment of MBR energy costs is some-
what higher than the early studies reported above.
Table 8 shows energy costs for the MBR itself.
Aeration costs for the microbiology are similar
to those of conventional activated sludge. To
this, aeration costs for fouling control need to be
added. Currently, an efficient system uses about
0.4 kW h/m3 for this purpose, though with system
design improvements, some systems may improve
on this level, and air lift systems appear to use
lower energy.
However, there are other energy costs for a
field system. The Singapore experience showed
an average use for commercial scale MBR of
1.0 kW h/m3. The Dutch consultancy DHV, who
have done extensive piloting of several MBRs
at Beverwijk in the Netherlands, are broadly in
Table 9
Energy usage for various water and wastewater sources
Feed Conv. act sludge
(kW h/m3)
Surface water
Wastewater 0.3–0.6
Wastewater MBR
Brackish (930–2200 ppm)
Brackish – Tidal Estuary
Seawater (medium salinity)
agreement with this figure, and suggest that practi-
cal experience indicates an energy use for MBR of
0.8–1.0 kW h/m3. However recent developments
are reducing air by progressively limiting wastage.
Consideration is also being given to anaerobic
MBRs, which would have lower energy usage.
The commercial application of both membrane
filtration and RO in the treatment of wastewater
feeds has shown that fouling rates can be signif-
icant at high fluxes. To minimize total water cost,
relatively low fluxes should be selected, increas-
ing capital costs whilst ensuring low operating
costs [8].
5. Summary of energy costs from
various sources
The cost of seawater RO (SWRO) varies
according to the salinity of the feed and the system
design. In this paper, energy costs have been taken
for a medium salinity feed (e.g. Mediterranean
Sea with a TDS of 38,000 ppm TDS) [9,10].
The summary of energy usage for the various
water and wastewater sources is summarized in
Table 9. For wastewater schemes, the conventional
activated sludge (CAS) process is required for
river discharge. In most cases, this level of treat-
ment is suitable for direct supply to an MF-RO
system without further biological treatment.
Therefore, the energy cost of CAS does not nec-
essarily need to be added to the treatment costs,
Pre-treatment RO system Total treatment
(kW h/m3) (kW h/m3) (kW h/m3)
0.15–0.3
0.1–0.2 0.4–0.5 0.8–1.0 (1.3)
0.8–1.0 0.4–0.5 1.2–1.5
0.1–0.3 0.6–0.9 0.8–1.0
0.29 1.38 1.67
0.3–1.0 2.0–3.0 2.3–4.0
 G.K. Pearce / Desalination 222 (2008) 66–73
depending on whether CAS treatment has to be
provided anyway.
The data for MBR has been based on practical
field experience. The wastewater figure assumes
CAS has to be provided also, but this level of
treatment is normally required irrespective of the
reuse scheme as noted above. Membrane filtration
and RO is added to the CAS to provide reuse
quality product. The figures for brackish are from
the large element consortium study in the US for
a 930 ppm surface water and a 2200 ppm ground-
water [3], with different pre-treatment require-
ments. Finally, the seawater scheme is based
on a medium salinity challenge. Pre-treatment
costs can become relatively high, depending on
recovery.
It should be noted that membrane filtration
and RO energy costs are relatively low in waste-
water reuse applications since it has been found
that these applications need to be operated at low
fluxes to avoid fouling. This creates low energy
cost, but the capex of the membrane system cost,
and opex of the membrane replacement costs are
higher than for similar salinity brackish water
treatment.
6. Conclusions
• Wastewater reuse is a well established appli-
cation and a rapidly growing market
• It is mandatory for a potable reuse scheme that
RO is used to provide the final treated water
quality, and that pre-treatment is provided by
membrane filtration
• The total energy cost of wastewater reuse
including the conventional activated sludge
stage followed by MF(or UF)-RO is 0.8–1.2
kW h/m3
• This compares with conventional surface
water treatment at 0.15–0.3 kW h/m3, brackish
water at 0.8–1.7 kW h/m3, and desalination
at 2.3–4.0 kW h/m3, thus reuse is relatively
73
attractive compared to other non-conventional
options
• MBR-RO schemes have an energy use similar
to brackish water at 1.2–1.5 kW h/m3, and
could be attractive for retrofit and Greenfield
site opportunities where space is at a premium
• Reuse schemes require mild operating condi-
tions and low membrane filtration and RO
fluxes to minimize power and chemical costs,
and avoid fouling problems
References
[1] J. Kolodziejski and C. Gasson, Water Reuse
Markets 2005–2015: A Global Assessment and
Forecast Global Water Intelligence, June 2005.
[2] 19th IDA Worldwide Desalting Plant Inventory:
Global Water Intelligence and Water Desalination
Report, June 2006.
[3] Industry Consortium Analysis of Large Element
RO/NF, Desalination & Water Purification Dev
Report 03-FC-81-0916, September 2004.
[4] R.D. Reardon, S.V. Paranjape, X.J. Foussereau,
F.A. DiGiano, M.D. Aitken, J.H. Kim, S.-Y. Chong
and R. Cramer, Water for Reuse: Membrane Treat-
ment of 2ry Effluent for Subsequent Use WERF
Final Report, North Carolina University, 2005.
[5] S. Adham, M. Kumar and W.H. Pearce, Model
developed for brackish and reclaimed water, Desal.
Water Reuse, 15 (3) (2005) 38.
[6] Energy Efficiency in Wastewater Treatment Waste-
water Engineering – Treatment and Reuse, 4th
edn., Metcalfe and Eddy, 2003, pp. 1703–1707.
[7] F. Rogalla, The Power to Cool Water & Waste-
water Treatment, 2006, pp. 41–43.
[8] Membrane Fouling Guidelines for Water and
Wastewater Treatment by UF/MF, in: Proceedings
of the IWA Conference, Harrogate, UK, May 2007.
[9] M. Wilf, The Guidebook to Membrane Desalination
Technology, Balaban Desalination Publications
(ISBN 0-86689-065-3), 2007.
[10] G.K. Pearce, C. Bartels and M. Wilf, Improving
total water cost of desalination by membrane
pre-treatment, in: Proceedings of the IDA Confer-
ence, Gran Canaria, October 2007 (in preparation).